my prefect server setup
prefect-metrics.waow.tech
python
orchestration
fix cache policy on run_curation_agent task
OpenAI/TurboPuffer client objects contain thread locks that can't be
serialized by Prefect's default cache policy. Set NONE explicitly.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+2
-1
+2
-1
flows/curate.py
+2
-1
flows/curate.py
···
20
20
from pydantic_ai.providers.anthropic import AnthropicProvider
21
21
from prefect import flow, get_run_logger, task
22
22
from prefect.blocks.system import Secret
23
+
from prefect.cache_policies import NONE
23
24
from prefect.variables import Variable
24
25
25
26
PHI_DID = "did:plc:65sucjiel52gefhcdcypynsr"
···
479
480
}
480
481
481
482
482
-
@task
483
+
@task(cache_policy=NONE)
483
484
async def run_curation_agent(
484
485
state_text: str,
485
486
session: dict[str, Any],